Hierarchical Line-to-Line Voltage-Based FCS-MPC Computational Cost Reduction for Multi-Level Converters

Finite Control-Set Model Predictive Control (FCS- MPC) offers a simple concept of control for multilevel converters (MLIs), ease of implementation, and the ability to address multiple control objectives while including constraints. However, a high computational burden is required by FCS-MPC, which further increases with additional voltage levels. This paper proposes a two-stage hierarchical FCS-MPC control strategy based on line-to-line voltage-based prediction model, splitting the control into current control and capacitor voltage balance, using switching state redundancies. The aim is to reduce the number of voltage level iterations by means of pre-defined line-to-line voltage level tables, thus the computational burden with fast dynamic response too. Furthermore, the strategy avoids the use of weighting factors. The proposed control strategy is validated on a three-cell flying capacitor converter with 512 switching states, obtaining a reduction of about 91% in iteration cycles, maintaining the dynamic performance of standard FCS-MPC.